RESOURCES
SYS.INT documentation covers authentication, deployment, and the v1 inference API in one place. Every guide ships a copy-paste code sample tested against the current API version.
Where do I start with SYS.INT?
Pick a card below: documentation, the API reference, a quickstart, the changelog, or live status. Every card jumps to a section on this page.
Documentation
Setup, authentication, and deployment guides for the SYS.INT platform.
API Reference
Every endpoint, request field, and response shape in the v1 inference API.
Quickstart Guides
Route your first inference request in one curl call, no SDK required.
Changelog
Dated release notes for every API, routing, and dashboard update.
Status
Current status for the inference API, edge routing, dashboard, and webhooks.
What does the SYS.INT documentation cover?
Four topics: authentication, deployment, rate limits, and webhooks. Each one below is a short, self-contained answer, not a link to a page that doesn't exist yet.
Authentication
Generate a key from the Contact flow and pass it as a Bearer token on every request.
Deployment
Deploy a fine-tuned model to the edge network with a single POST request.
Rate limits
Open Source keys are capped at 10,000 requests per month. Paid tiers are unlimited.
Webhooks
Subscribe to deployment and inference events over a signed HTTPS callback.
What endpoints does the SYS.INT API expose?
Five endpoints under /v1 handle routing, models, deployment, usage, and status. Every request and response is JSON over HTTPS.
| Spec | DESCRIPTION |
|---|---|
| POST /v1/route | Route an inference request to the lowest-latency edge node. |
| GET /v1/models | List all 147 foundation models available to the current key. |
| POST /v1/deploy | Deploy a fine-tuned model to the edge network. |
| GET /v1/usage | Return request counts and latency percentiles for the billing period. |
| GET /v1/status | Return current health for the inference API and edge network. |
Every request needs an Authorization: Bearer header. See the quickstart below for a full working example.
How do I send my first request?
Send one curl call with your API key and a prompt. The example below routes to the sysint-core-14b model and returns a response in about 12ms, the same p99 latency shown on the homepage.
1. Get a key
Request an API key through the Contact page. Open Source keys are free.
2. Call the endpoint
POST to /v1/route with your prompt. No SDK install required.
3. Read the response
The response includes the model output plus the routed region and latency.
$ curl https://api.sysint.template/v1/route \
-H "Authorization: Bearer $SYSINT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "sysint-core-14b",
"input": "Summarize the attached incident report.",
"max_tokens": 512
}'// RESPONSE (12ms)
{
"id": "req_8f2c1a90",
"model": "sysint-core-14b",
"region": "us-east-1",
"latency_ms": 12,
"output": "..."
}Is the SYS.INT API online right now?
All four services below are operational in this demo. The table is representative status data for the template, not a live monitoring feed.
| Spec | STATUS |
|---|---|
| Inference API | Operational |
| Edge Routing | Operational |
| Dashboard | Operational |
| Webhooks | Degraded |
What changed in the most recent SYS.INT release?
v1.3.0 shipped on 2026-08-18: latency-aware edge routing. The releases before it are listed below with their dates.
Latency-aware edge routing
Edge routing now picks the nearest of 50+ regions by measured latency instead of static geography. Median routing overhead dropped from 3ms to 1.4ms in internal testing.
Usage report fix
Fixed a bug where GET /v1/usage double-counted retried requests in the billing period total.
Usage endpoint
Added GET /v1/usage, returning per-key request counts and latency percentiles.
Signed webhooks
Webhook payloads now include a signature header so receivers can verify the request origin.
Questions about SYS.INT resources
01Is the SYS.INT documentation free to read?
Yes. Documentation, the API reference, and the quickstart guide on this page are free and do not require an account.
02Do I need an API key to run the quickstart example?
Yes. Request a key from the Contact page first. The Open Source tier includes 10,000 requests per month at no cost.
03Where can I check current API uptime?
The Status section on this page lists current status for the inference API, edge routing, dashboard, and webhooks, plus the 99.97% uptime figure quoted on the Pricing page.
04Is there an official SDK?
Not yet. Call the REST API directly with curl, fetch, or any HTTP client. A typed SDK is on the roadmap.
START ROUTING REQUESTS
Request access through the Contact page. No credit card required for the Open Source tier.
GET API ACCESS